Weiyao Wang dedicated eight years to Meta — his inaugural position after college — assisting in the development of multimodal perception frameworks and taking part in open-world segmentation initiatives, including SAM3D. His last day at Meta occurred last week, and he has now transitioned to Thinking Machines Lab (TML).

Wang’s transition to TML coincides with the AI startup’s expansion across various sectors. It recently secured a multibillion-dollar cloud agreement with Google, granting it access to Nvidia’s newest GB300 chips and positioning it as one of the initial startups to utilize this hardware.

The contract, revealed this past Tuesday at Google Cloud Next, follows a prior collaboration with Nvidia, placing TML alongside Anthropic and Meta in the equivalent infrastructure category. (Meta reportedly considered acquiring Thinking Machines around the same time last year and has recently been recruiting TML’s founders individually.)

The landscape of talent is shifting. Wang and Kenneth Li — a Harvard PhD who spent 10 months at Meta before joining TML this month — exemplify a two-way talent exchange. Business Insider indicated last week that Meta has now recruited seven of TML’s founding members. An examination of recent hires suggests Thinking Machines is also actively recruiting from Meta. Based on a review of LinkedIn profiles, TML seems to have attracted more researchers from Meta than from any other single organization.

Among the most notable is Soumith Chintala, TML’s CTO, who spent 11 years at Meta and co-founded PyTorch, the widely-used open source deep learning framework that forms the foundation for much of today’s AI research. He departed Meta in late 2025 and took on the CTO role earlier this year. Piotr Dollár, another 11-year veteran of Meta who served as research director and co-authored the seminal Segment Anything model, is now part of TML’s technical team. Andrea Madotto, a research scientist in Meta’s FAIR division with a focus on multimodal language models, joined TML in December. James Sun, a software engineer with almost nine years at Meta concentrating on LLM pre- and post-training, also made the switch.

TML has also recruited talent from outside Meta. Neal Wu — a three-time gold medalist at the International Olympiad in Informatics and a founding member of the popular coding startup Cognition — came on board earlier this year. Jeffrey Tao transitioned from Waymo, Windsurf, and OpenAI. Muhammad Maaz previously held a research fellowship at Anthropic. Erik Wijmans joined from Apple. Liliang Ren spent two and a half years with Microsoft’s AI Superintelligence group pre-training OpenAI models for code before coming on board in March.

The startup currently employs around 140 people.

Meta’s compensation packages — reaching seven figures without conditions — are now widely recognized. For researchers assessing their alternatives, the equation might be straightforward: Thinking Machines Lab is presently valued at $12 billion. While this amount may seem extraordinary for a company at this stage during any prior tech cycle (having launched only one product thus far), in comparison to the record valuations of OpenAI and Anthropic, the financial potential remains substantial.

Contacted on Friday morning, a representative from TML opted not to provide comments for this article.